1. Field of the Invention
The present invention relates to a mode adaptation method for a Multi-Input Multi-Output (MIMO) system, and more particularly to a method for reducing an amount of uplink feedback information.
2. Discussion of the Related Art
Closed-loop MIMO (CL-MIMO) systems, which are adaptively operated according to a channel condition of a user, among all MIMO systems based on a multi-antenna can greatly improve a performance or throughput of an overall system.
The closed-loop MIMO systems can be classified into a codebook based preceding system for reporting a preceding matrix index (PMI) and an analog feedback based preceding system for reporting a channel. The codebook based precoding system and the analog feedback based preceding system may have the following advantages and disadvantages.
The codebook based preceding system requires a small amount of feedback information, has a small number of channel quality indicator or information (CQI) mismatches, and reduces a flashlight effect according to codeword restriction. At this time, the CQI may be measured in either a midamble or common pilots on the basis of complete comprehension of both a precoder and an interferer.
The codebook based preceding system has the following disadvantages. In more detail, the codebook based preceding system has a large number of quantization errors and a low beamforming gain. In order to implement the above-mentioned codebook based preceding system, an appropriate codebook design is needed.
The analog feedback based preceding system reduces the number of quantization errors and acquires a high beamforming gain. On the other hand, the analog feedback based preceding system requires an excessive amount of either feedback information or feedback errors, has a large number of CQI mismatches, and generates the flashlight effect. Specifically, if the analog feedback based preceding is used for a multi-user MIMO system, the CQI can be measured in only dedicated pilots. Also, provided that the analog feedback based preceding system does not recognize the interfering precoder, the mobile station is unable to measure the CQI. In order to implement the analog feedback based preceding system, a sounding channel is needed.
Therefore, the codebook based preceding system has the above-mentioned advantages, such that it has been widely used for a 3GPP LTE system, a WiMAX system, and a 2GPP2 Ultra Mobile Broadband (UMB) system.
The MIMO systems are classified into a single user MIMO (SU-MIMO) system and a multi-user MIMO (MU-MIMO) system according to methods for allocating the spatial resources.
FIGS. 1 and 2 illustrate block diagrams illustrating transmission structures when data is transferred at two or more spatial multiplexing rates.
FIG. 1 shows a case in which a vertical encoding (or a single codeword (SCW)) are used. FIG. 2 shows another case in which horizontal encoding (or a multi-codeword (MCW)) are used.
The SU-MIMO system allocates all of the spatial resources to only one mobile station (MS). When the SU-MIMO system is operated under a closed-loop MIMO scheme, each mobile station (MS) selects a preferred rank (i.e., a spatial multiplexing rate), and reports the preferred rank, a preceding matrix index (PMI) and a channel quality information (CQI) which are suitable for the selected rank. A base station (BS) allocates only one mobile station (MS) to a resource (i.e., time and frequency) using such feedback information. In this case, the spatial resources are all used by the mobile station (MS). Here, the SCW or MCW may be used as the transmission structure.
The MU-MIMO system is used to allocate spatial resources to a few mobile stations. When the MU-MIMO system is operated under the CL-MIMO scheme, each mobile station (MS) transmits the CQI and the PMI according to MU-MIMO conditions. In this case, the MU-MIMO conditions are a precoding matrix set, MU-MIMO types (PU2RC, ZF-BF, . . . ), and the like. The base station (BS) selects mobile stations (MSs) satisfying a specific condition using the received information, and allocates the selected mobile stations to a resource (i.e., time and frequency). Here, the transmission structure is set to the MCW.
Generally, if the number of users is small, the SU-MIMO system has a good throughput superior to that of the MU-MIMO system. If the number of users is large, the MU-MIMO system has a good throughput superior to that of the SU-MIMO system. When several users are paired with each other in the MU-MIMO system, if orthogonal pairs of users are found, the throughput of the MU-MIMO system becomes better. The larger the number of users, the higher the probability of generating the orthogonal pairs.
FIG. 3 is a graph showing a comparison of throughput between the SU-MIMO system and the MU-MIMO system.
In FIG. 3, if about 5 or more people are in a cell, the MU-MIMO system has a good throughput superior to that of the SU-MIMO system. This specific point is called an SU-MU switching point. This switching point may occur at different positions according to a channel status, a user status and the like.
In more detail, a throughput of an optimum system must follow envelopes of the SU-MIMO throughput and the MU-MIMO throughput. For this operation, the mobile station (MS) must carry out not only a feedback operation suitable for the SU-MIMO system but also another feedback operation suitable for the MU-MIMO system. As a result, an amount of feedback overhead unavoidably increases.
Generally, a codebook structure is pre-decided. Specifically, in case of the MU-MIMO mode, if the number of users is not very large, a codebook size must be small such that a pairing gain arises. Therefore, the codebook size for the MU-MIMO mode has generally been designed to be smaller than that of the SU-MIMO mode, such that a beamforming gain decreases.